View Single Post
Old 04-12-2008, 10:52 AM   #41 (permalink)
SteveP
EcoModding Lurker
 
Join Date: Apr 2008
Location: SF Bay Area
Posts: 35
Thanks: 0
Thanked 3 Times in 3 Posts
Quote:
Originally Posted by cbergeron View Post
He once asked his students to determine the volume of a lightbulb and they scribed formulas on their notepads for hours upon hours on end. Eventually a solution was reached on paper. To wit - Edison took a lightbulb, unscrewed it and poured water into it until it was full. He then, poured the bulb full of water into a flask and measured the volume of a lightbulb - not by measure, but in practice (experience); and in a matter of seconds.
Ah, but the only reason that approach worked is that the lightbulb whose volume was to be measured already existed! Suppose, on the other hand, it would take $1000 and six weeks to make the lightbulb whose volume you cared about--mightn't it be worth a a few hours of scratchpad time to calculate it in advance, especially if the volume was critical to some other aspect of the project (like fitting (volume-wise) in some sort of enclosure)? Suppose that first bulb turned out to be too big? Now you have to spend another $1000 and six weeks to make a smaller one to measure. [I know, you could build a *model* of a lightbulb -- cheaper and faster-- and measure that, but I hope my point is clear.]

On the other hand, Edison was forced to find filament material by trial and error--the theory of what makes a good filament didn't exist then (and maybe not even now, for all I know). Sometimes it pays to do some research and calculatin'. Sometimes it doesn't. Most of the time you have to do both.

Remember, this discussion started when the OP *did* the experiment (with the air dam). It didn't (appear to) work. OK, what next? Decide that "air dams don't work"? Look into the measurement methods that were used in case those were faulty? Give up? Figure out what the variables are and see which ones might be worth adjusting to get better results? Try things purely at random and hope that with luck something proves beneficial? How long would it take to run, say, a dozen experiments, making sure weather conditions (temp, wind, humidity) were nearly identical for each experiment?

Experiments always *look* easy--until you dig into the measurement issues, the number of variables that are involved, the time to build the test objects, and so on. It always *looks* like it will only take an experiment or two to decide the issue--but it usually takes a lot more simply because the early results are often contradictory, or at odds with current understanding, and so on. Take the air dam that started this thread--ignoring measurement issues, we could: change the clearance between the air dam and the road, say 0.5" per test. We could put a "chin" of various dimensions on it. We could narrow its height in the center but leave it "tall" in front of the wheels, and do that experiment 0.5" at a time. And, of course, we could change the ride height. But what if it's the case that it takes a combination of things to make the air dam effective--now we have to test combinations of changes and that really increases the number of experiments we'd have to do.

Heck, if someone *gave* me a wind tunnel (i.e., accurate testing was relatively cheap), I *still* would read the literature before I started testing cars. And I'm saying that as someone who unfortunately has to read the darned SAE papers on crummy microfiche at a library that's 30 miles away. (Yes, the irony is there: I'm burning gas in order to hopefully save some in the future )

That said, I'm not against experimentation. Really

--Steve

P.S. LostCause: I'm not overlooking your comment--please let us know what you find out.

P.P.S. I don't want to sound too discouraging or anything--what we're trying to achieve is *hard*. Aero mods (generally) aren't silver bullets where one mod will buy us gobs of efficiency. Rather, it's a series of small mods which, when added up, buy us a significant improvement. Speaking as someone who's done a bit of calculation, we're talking about changes which individually are on the order of a few percent. Those are tough to measure outside of a lab setting.


Last edited by SteveP; 04-12-2008 at 11:02 AM.. Reason: Added PPS
  Reply With Quote